Effects of aerated and non-aerated biofilters on effluent water treatment from a small-scale recirculating aquaculture system for Nile tilapia (Oreochromis niloticus L.)

The global capture fisheries production has stagnated since 1980s, and in 2016, the total capture fisheries production was 90.9 million tons, representing a decrease of almost 2 million tons from 92.7 million tons in 2015. At the same time, there has been an increasing global demand for fish (FAO, 2018). FAO estimates indicate an average annual increase in global fish consumption of 3.2% between 1961 and 2016 (FAO, 2018). The expansion of aquaculture sector will probably provide a solution to the stagnating capture fisheries and meet the demand for fish (Waite et al., 2014). However, common aquaculture production systems such as ponds may not contribute significantly to aquaculture production because they require more land as well as large volumes of water for expansion (Verdegem et al., 2006). Therefore, to meet the increasing global demand for fish, aquaculture production must be intensified to minimize reliance on limited land and water resources (Waite et al., 2014). However, pond-based systems allow limited options for intensification. Besides, the systems are operated in a flow-through system where water is frequently exchanged with fresh water from adjacent rivers to maintain water quality for fish production (Liti et al., 2005). The discharge of untreated aquaculture waste-water contaminates the quality of water of the receiving aquatic ecosystems (Ramírez-Godínez et al., 2013; Turcios and Papenbrock, 2014).

Recirculating aquaculture systems are recognized as suitable systems that can increase production, minimize wastewater discharge as well as reliance on limited land and freshwater resources (Badiola et al., 2012). Removal of waste in these systems is generally managed through mechanical removal of solids and conversion of ammonia and nitrite to nitrate through the nitrification process (Hamlin et al., 2008). But, nitrification does not result in the overall reduction of nitrogenous wastes but merely transforms ammonia to nitrate, resulting in nitrate accumulation and lowering of the pH of culture water (Piedrahita, 2003). Therefore, RAS requires 10–20% of water exchange each day to reduce accumulation of nitrates and maintain suitable water quality for fish (Martins et al., 2010). Such water exchange increases water consumption and the cost of fish production (Meriac, 2014). Besides, water exchange is undesirable due to the potential negative impact on the cultured fish and the recipient environment (Davidson et al., 2009). Davidson et al. (2014) demonstrated that high nitrate concentrations in RAS with minimal water exchange can cause chronic toxicity to fish. Further, the nitrifying bacteria have long start-up and multiplication periods and require high oxygen concentrations which translates to higher energy costs (Delong and Losordo, 2012; Cristian et al., 2013). The bacteria are also sensitive to rapid changes in pH, temperature and water flow and take longer to recover in the event of system failure (Lee et al., 2000).

Furthermore, one of the major constraints to the widespread of RAS, particularly in the Sub-Saharan African countries such as Kenya, is the high energy costs connected to its operation (Opiyo et al., 2018). Hence, there is a need for systems that can lower energy costs, improve productivity and efficiency. For this reason, current technical innovations in RAS are searching for more energy and cost-efficient systems such as the incorporation of denitrification biofilter in freshwater RAS to reduce the operating costs by 10% (Martins et al., 2010). Denitrification reduces inorganic nitrogen compounds such as nitrite and nitrate to elemental nitrogen and can provide solutions to the limitations of nitrification (Gichana at al., 2018). Facultative anoxic bacteria fuel the process with electron donors from either organic or inorganic sources (van Rijn et al., 2006). The anoxic bacteria are most abundant in the aquatic environment (Dalsgaard et al., 2003) and have higher growth rates of 4.8 day⁻¹ than that of the nitrifying bacteria having growth rate of 0.76–0.84 day⁻¹ (Fdz-Polanco et al., 2000). Moreover, denitrification reduces water exchange requirements and power consumption (van Rijn et al., 2006). Integrating denitrification in RAS reduces waste discharge by 81% for nitrogen, 59% for chemical oxygen demand, 61% for total oxygen demand, 58% for total solids and 30% for carbon dioxide (Martins et al., 2010). Despite the considerable advantages of introducing denitrification in recirculating systems, its use in commercial aquaculture is still limited and the aquaculture industry has been slow to adopt this technology. Major reasons include: (1) the requirement to circulate water through aerobic biofilters, (2) the need for skilled personnel for operation and maintenance, and (3) the alternative use of a costly external carbon source (van Rijn et al., 2006; Martins et al., 2010).

Studies evaluating biofilter performances are becoming more common in RAS, but most of these studies are done at laboratory or on pilot scale. Assessment of biofilters in the actual culture conditions is urgently needed (Guerdat, 2008). Furthermore, most denitrification studies use nitrifying biofilter before the denitrifying biofilter to convert ammonia to nitrates, which is thereafter converted into elemental nitrogen in the denitrification chamber. Studies incorporating nitrification and denitrification processes in a single biofilter are few or non-existent. The aim of our study was to: (1) design a single biofilter that requires very low maintenance and incorporates nitrification and denitrification processes, (2) compare the biofiltration efficiency of the non-conventional non-aerated biofilter and the conventional aerated biofilter and (3) evaluate the effect of the two biofilters on the growth of O. niloticus. This study was motivated by the fact that different nitrogen cycling bacteria including nitrifiers and denitrifiers coexist in a biofilm as long as conditions that favor their growth are created (Egli et al., 2003; van Kessel et al., 2010).

Water temperature in the RAS was within acceptable ranges for the growth of nitrifiers (Chen et al., 2006), denitrifiers (Wang and Chu, 2016) and tilapia (Guerdat et al., 2010). The average pH in the two biofilters was also within the optimum range for the biofiltration process (6.7–9.0) (Chen et al., 2006), the survival and growth of tilapia (Delong et al., 2009). In any aquaculture system, dissolved oxygen is one of the most critical parameters that maintain safe limits and provide optimal growth conditions for the fish (Timmons et al., 2002). Mean dissolved oxygen was lower than preferred 5 mg L⁻¹ for optimum growth of tilapia (Colt, 2006) and within acceptable limits for nitrification (> 1.0 mg L⁻¹) (Timmons and Ebeling, 2010). But tilapia can tolerate oxygen concentrations as low as 1.0 mg L⁻¹, and below this level, fish may utilize atmospheric oxygen from the surface film of water (Ross, 2000). Oxygen is also one of the major factors that affect the rate of nitrification in aerated biofilters because nitrifiers are obligate aerobes, which require oxygen to survive and multiply (Timmons and Ebeling, 2010). Low oxygen levels decrease the rate of nitrification, though levels as low as 2 mg L⁻¹ are sufficient to maintain nitrification, while levels below 2 mg L⁻¹ inhibits the process (Timmons and Ebeling, 2010). Non-aerated biofilters depend on facultative anaerobes to convert ammonia or nitrates and nitrites to dinitrogen gas (van Rijn et al., 2006). Therefore, DO of approximately 0.5 mg L⁻¹ or less can allow non-aerated biofiltration in both freshwater and marine systems (van Rijn et al., 2006). Conversely, denitrification can occur in aerated conditions with DO levels as high as 4–5 mg L⁻¹. However, the rate of denitrification decreases in aerated conditions (Gutierrez-Wing et al., 2012).

The high ammonia levels at the beginning of the study suggest that biofilters had not appropriately established. Delong and Losordo (2012) stated that the bacteria responsible for biofiltration take an extended period to grow and multiply. Our findings are consistent with Wongkiew et al. (2017), who reported a build-up of ammonia concentrations during the start-up periods in a recirculating aquaculture system. The consistent low and high nitrate concentration in the non-aerated and aerated biofilter outlets is an indication of nitrate removal and production through denitrification and nitrification processes. High nitrite concentration in the aerated biofilters indicates a higher rate of ammonia oxidation than nitrite oxidation. The high rate of oxidation is due to the shorter generation times of ammonia oxidizing bacteria which increases the growth of more ammonia oxidizing bacteria compared to nitrite oxidizing bacteria and subsequent accumulation of nitrite (Wongkiew et al., 2017). Accumulation of nitrites in the non-aerated biofilters can be attributed to incomplete denitrification process, which reduces nitrates to nitrites with no further reduction (Stief, 2001). Ramdhani and Bux (2007) showed that incomplete denitrifiers dominate the denitrifying biofilter and are therefore significantly involved in nitrogen removal during wastewater treatment.

The difference in ammonia removal efficiency between aerated and non-aerated biofilters can be an indication of different biological processes that occurred in the biofilters because of different environmental conditions. Ammonia removal is a function of the ammonia oxidizing bacteria in aerobic and anaerobic conditions (Malone and Pfeiffer, 2006). Under aerobic conditions, ammonia-oxidizing bacteria converts ammonia to nitrates resulting in low ammonia and high nitrate levels in aerated biofilters. Whereas, under anaerobic conditions, anaerobic oxidation of ammonia (anammox) oxidizes ammonia to molecular nitrogen (van Kessel et al., 2010). This can explain the removal of ammonia in the non-aerated biofilter. The ammonia removal efficiency of 37.7% is within the range (11–80%) reported by previous studies using nitrifying or nitrifying-denitrifying biofilters in RAS (Kir, 2009; Nootong and Powtongsook, 2012; Tsukuda et al., 2015). Besides, fine solid accumulation and low oxygen levels in the non-aerated biofilter probably favored the denitrification process because low oxygen levels make denitrifying bacteria more competitive than nitrifying bacteria (Bartelme et al., 2017). Denitrification reduces oxidized inorganic nitrogen compounds such as nitrites and nitrates to elemental nitrogen under anaerobic and anoxic conditions using carbon as an electron donor (van Rijn et al., 2006). Therefore, the denitrification process contributed to the significant removal of nitrates in the non-aerated biofilter. Moreover, the significant removal of phosphorus in the non-aerated biofilter can be attributed to denitrification. Studies have shown that some denitrifiers can uptake more phosphorus than their metabolic requirements resulting in low phosphorus in the bioreactor (Krom et al., 2014).

The fish showed poor feed conversion efficiency in both treatments, most probably due to high ammonia concentration. The relatively high ammonia concentration in the culture system might have reduced fish appetite, which reduced feed intake and increased FCR values (Effendi et al., 2017). The mean FCR values were higher than the recommended values of 1.5–2.0 for intensively reared tilapia (Stickney, 2005). The specific growth rates were lower than those from previous studies, which reported growth rates in the range of 0.86% day⁻¹ to 3.09% day⁻¹ in RAS with either aerated and non-aerated biofilters (Shnel et al., 2002; Al-Hafedh and Alam, 2007; Arredondo-Figueroa et al., 2015). The low growth rates in this study were probably due to the long exposure of fish to ammonia concentrations above the acceptable limits of 1.0 mg L⁻¹ (Delong et al., 2009). Furthermore, ammonia and nitrite concentrations had a negative relationship with the weight of fish, suggesting that for every increase in one unit of ammonia and nitrites, there was a corresponding decrease in the growth rate of fish. The survival rates were lower than 97% obtained by Nootong and Powtongsook (2012) with tilapia grown in a recirculating aquaculture system with submerged fibrous nitrifying biofilters. The final bio-mass obtained in this experiment was within the range of 12.35–30.35 kg m⁻³ reported by Nootong and Powtongsook (2012). The findings indicate that the non-aerated biofilters did not negatively affect the growth of fish.

The aerated and non-aerated biofilters reduced wastes generated from the recirculating aquaculture system. Most water quality parameters were within ranges, considered appropriate for fish and biofiltration processes. Although the present study did not use aerated biofilters before the non-aerated biofilters, the latter was able to remove ammonia from the system. Our results demonstrated the feasibility of using non-aerated biofilters to maintain water quality for the production of O. niloticus and reduce the discharge of nitrogen and phosphorus wastes to the surrounding environment. Besides, non-aerated biofilters are as effective as aerated ones, suggesting that the non-aerated biofilters can provide opportunities to run the system without the need to reactivate the biofilters in case of system failure. The results also suggest that there is potential to fine-tune biofilter design to take advantage of the coexistence of the nitrification and denitrification bacteria in the biofilters. However, further research is required to fully understand the implications of integrating nitrification and denitrification processes in a single biofilter on nitrogen removal in aquaculture systems.